Token导航 LogoToken导航

Agent Skill · 待分类

mdp(平均发展计划)

mdp 用于处理 GitHub 仓库、Issue、Pull Request 和代码协作信息,适合在 Codex、Claude、Cursor、Gemini CLI 中需要围绕仓库状态、代码变更或协作事项进行整理时使用。可结合来源仓库、安装命令和原始 README 继续核验具体用法。安装前建议确认权限范围、维护状态,以及是否会触发联网、命令执行或文件读写。

mdp 是待分类类 Agent Skill,可用宿主为 Codex、Claude、Cursor、Gemini CLI,通过GitHub 仓库获取,权限为只读;许可证信息还不明确,第三方扫描有未通过项,按安装量在待分类类的 11,209 个 Skill 里排第 9,754。

总安装 225周安装 9GitHub Stars 公开资料未说明下载 73核验 2026-05-01

安装说明

本站只整理中文说明,不托管安装包,也不代用户安装。
GitHub
许可证
unknown

安装方式

通过对话安装

复制提示词发给支持本地命令或 Skills 的 AI 助手,先确认命令和权限,再让它执行。

请帮我安装这个 Agent Skill:mdp(平均发展计划)
来源仓库:https://github.com/varunpandey0502/markdown-projects
仓库路径:skills/mdp
安装命令:
npx skills add https://github.com/varunpandey0502/markdown-projects --skill mdp
安装前请先检查当前环境是否支持对应 CLI,并向我确认将要执行的命令、安装目录、联网范围和文件读写权限;确认后再执行。

命令行安装

复制命令到本机终端执行。该命令会通过 npx skills 从第三方来源获取 Skill;本站只展示命令,不托管安装包,也不自动执行。

skills.shnpx skills
npx skills add https://github.com/varunpandey0502/markdown-projects --skill mdp

这个 Skill 是做什么的

这个 Skill 的分类还没最终确定,用法以来源仓库的说明为准。

mdp 用于处理 GitHub 仓库、Issue、Pull Request 和代码协作信息,适合围绕仓库状态与协作事项进行整理。适用于待分类内容的协作信息管理场景。通过 npx skills add 命令从指定 GitHub 仓库安装并使用。安装前需确认权限范围、维护状态及是否触发联网或文件读写操作。

mdp 属于待分类类 Skill,可作为该场景下的辅助能力补充。

来源仓库内路径:skills/mdp

上手步骤

4 步
  1. 1

    确认自己的宿主工具

    公开资料标注支持 Codex、Claude、Cursor、Gemini CLI;先确认自己用的是其中之一,再按对应工具的 Skills 目录放置。

  2. 2

    拿到安装命令

    可以直接复制的命令是 npx skills add https://github.com/varunpandey0502/markdown-projects --skill mdp;命令里的占位符换成自己的路径或密钥再执行。

  3. 3

    安装后重启宿主

    命令执行完要完全退出宿主工具再打开,让新装的 Skill 被重新扫描到,否则列表里可能看不到。

  4. 4

    按最小权限试用

    它标注为只读,先用只读任务试一次,确认行为符合预期再扩大使用范围。

常见问题

6 条

mdp 怎么安装?

在终端执行 npx skills add https://github.com/varunpandey0502/markdown-projects --skill mdp。装完完全退出宿主工具再打开,让它重新扫描本地 Skill。

mdp 支持哪些 AI 工具?

公开资料标注支持 Codex、Claude、Cursor、Gemini CLI。其他工具只要能读取本地 Skill 目录,一般也能手动放进去,但要自己确认触发方式是否一致。

mdp 需要什么权限?

权限级别是只读。该 Skill 主要提供规则、说明或参考内容,本身偏只读;真正读写文件、联网或执行命令仍取决于宿主 Agent 的任务。

mdp 是免费的吗?

公开资料没有标注明确许可证,收费和商用限制都无法确认。正式使用前请到来源仓库核对授权说明。

mdp 用的人多吗?

公开数据是累计安装 225 次,近一周安装 9 次,下载 73 次。这些数字来自第三方目录统计,只能当作热度参考,不代表质量或安全性。

mdp 安全吗?

第三方扫描有未通过项,风险较高。除非你确认过源码,否则不建议直接安装到常用环境。

值得选的地方和要先想清楚的地方

值得选的地方

  • 同时适配 4 种工具,换客户端不用重新找替代品。
  • 标注为只读权限,试用阶段不容易误改本地文件。

先想清楚的地方

  • 许可证没有明确标注,能不能商用、能不能再分发都不确定。
  • 第三方扫描存在提醒项或未通过项,需要人工逐条确认。
  • 第三方目录整理,版本更新和行为细节以来源仓库为准。

适不适合你

更适合

  • 已经在用 Codex,想直接扩展它能力的人。
  • 需要处理待分类类任务、又不想每次都手写提示词的人。
  • 只想先试试效果、不希望它改动本地文件的人。

不太适合

  • 需要明确授权才能用到公司项目里的场景。
  • 需要厂商级技术支持和 SLA 的生产系统。

待分类类 Skill 对比

6 个同类
Skill总安装GitHub 星权限许可证扫描
infsh-cli

infsh CLI · Codex、Claude

4,374,576382权限未标注未明确未收录
ai-video-generation

AI 视频生成 · Codex、Claude

4,313,376382调用外部服务未明确有提醒
lark-openapi-explorer

云雀 OpenAPI 浏览器 · Codex、Claude

2,064,1209,011需要联网未明确有提醒
lark-workflow-standup-report

Lark 工作流程站立报告 · Codex、Claude

2,041,6328,978需要联网未明确通过
use-my-browser

使用我的浏览器 · Codex、Claude

1,848,49140操作浏览器未明确未通过
mdp

平均发展计划 · 本页

225—只读未明确未通过

按安装量取同分类前几名;数字来自第三方目录统计,只作热度参考,不代表质量结论。

SKILL.md 原文

mdp — Markdown Project Management

A file-based project management CLI. Projects live in .mdp/ directories containing markdown files with YAML frontmatter.

IMPORTANT: On first use

When you start working with an existing project, immediately read .mdp/settings.json to learn the project's valid statuses, types, labels, and priorities, and read .mdp/project.md for project identity, health, and instructions. Do not guess or assume defaults — the config is the source of truth. Statuses are grouped by category (e.g., completed, started), and only the name field within each status is used in commands.

Installation

bun install -g github:varunpandey0502/markdown-projects

See INSTALL.md for details.

Quick start

mdp project create -p . --preset software
mdp issue create -p . -t "Fix login bug" --type bug --priority High
mdp issue list -p .
mdp issue update -p . --id ISS-1 -s "In Progress"

Project structure

.mdp/
├── settings.json          # Schema config: statuses, priorities, labels, types (committed)
├── project.md             # Project identity: title, description, instructions, health, log (committed)
├── issues/                # Flat directory of issue folders
│   └── ISS-1-add-auth/
│       └── ISS-1-add-auth.md
├── milestones/            # Flat directory of milestone folders
│   └── M-1-v1-release/
│       └── M-1-v1-release.md
├── docs/                  # Documentation
└── templates/             # Issue/milestone templates

Configuration

Schema config is stored in .mdp/settings.json with entity-scoped objects:

{
  "issues": {
    "prefix": "ISS",
    "statuses": {
      "triage": [], "backlog": [...], "unstarted": [...],
      "started": [...], "completed": [...], "canceled": []
    },
    "priorities": [...], "labels": [...], "types": [...]
  },
  "milestones": {
    "prefix": "M",
    "statuses": {
      "backlog": [], "planned": [...], "in_progress": [...],
      "completed": [...], "canceled": []
    },
    "priorities": [...], "labels": [...]
  }
}

Project identity is stored in .mdp/project.md with YAML frontmatter:

---
title: My Project
description: Optional one-line project description
instructions: Optional free-text guidance for LLMs and collaborators
health: on-track
log: []
createdAt: 2025-01-01T00:00:00.000Z
updatedAt: 2025-01-01T00:00:00.000Z
---

Statuses are grouped by status category (lifecycle stage). Each category maps to an array of {name, description} status objects. The system uses categories (not status names) to determine completion, overdue detection, etc.

Issues and milestones have independent statuses, priorities, and labels.

At runtime, .mdp/settings.json is the sole source of truth for schema config — no merge layers.

At project creation: preset (built-in or custom from ~/.mdp/settings.json) → CLI flag overrides → written to settings.json.

Custom presets and default preferences (default preset, output format) are stored in ~/.mdp/settings.json.

Global options

Every command accepts: -p <path> (project path), -f json|table (output format), -q (quiet), -V (verbose).

All output is JSON: {"ok": true, "data": {...}} or {"ok": false, "error": {...}}.

Commands

Project management

  • mdp project create -p <path> [--preset <name>] [-F|--force] [--with-templates] [--no-with-templates] [--issue-prefix <prefix>] [--milestone-prefix <prefix>] [--tags <tags>] [--title <title>] [--description <desc>] [--instructions <text>] — Create a new project (presets: software, marketing, design, product, social-media, generic)
  • mdp project get -p <path> [--no-include-content] — Get project identity, health, log, and body
  • mdp project settings -p <path> — Show project schema settings
  • mdp project stats -p <path> — Project statistics
  • mdp project fix -p <path> [--dry-run] — Fix folder structure to match frontmatter

Project log

Manage log entries on the project (stored in .mdp/project.md):

  • mdp project log add -p <path> -b "message" [--author <name>] [--health on-track|at-risk|off-track] [--dry-run]
  • mdp project log list -p <path>
  • mdp project log get -p <path> --index <n>
  • mdp project log update -p <path> --index <n> [-b <body>] [--author <name>] [--health <health>] [--dry-run]
  • mdp project log delete -p <path> --index <n> [--dry-run]

Project registry

Manage projects in ~/.mdp/settings.json:

  • mdp project list [--tag <tag>] — List registered projects (includes tagDescriptions in output)
  • mdp project add <path> [--tags <tags>]
  • mdp project remove <path>
  • mdp project tag <path> --add <tags> | --remove <tags> — Tags added via --add are auto-created globally

Tags

Manage tags globally via mdp tag:

  • mdp tag list — List all tags with descriptions and project counts
  • mdp tag add <tag> [-d "description"] — Create a tag (description defaults to "")
  • mdp tag update <tag> -d "description" — Update a tag's description
  • mdp tag remove <tag> [--force] — Remove a tag; --force to also strip from projects

Issues

Create
mdp issue create -p <path> -t "Title" [options]
OptionDescription
-t, --title <title>Issue title (required)
--type <type>Issue type (config-driven)
-s, --status <status>Initial status (default: first status in config)
--priority <priority>Priority level (default: null)
-l, --labels <labels>Comma-separated labels
-a, --assignee <assignee>Assignee identifier
-m, --milestone <milestone>Milestone ID
-e, --estimate <estimate>Effort points (positive integer)
--spent <spent>Actual effort spent
--due-date <date>Due date (YYYY-MM-DD)
--blocked-by <ids>Comma-separated issue IDs
--parent <id>Parent issue ID
--related-to <ids>Comma-separated issue IDs
--checklist <items>Comma-separated checklist items
-d, --description <desc>Short description
-c, --content <content>Full markdown body (or - for stdin)
--template <name>Template name from.mdp/templates/
--dry-runPreview without creating
List
mdp issue list -p <path> [options]
OptionDescription
-s, --status <statuses>Comma-separated status filter
--type <types>Comma-separated type filter
--priority <priority>Priority filter
-l, --labels <labels>Comma-separated labels filter
-a, --assignee <assignee>Filter by assignee (none for unassigned)
-m, --milestone <milestone>Filter by milestone ID (none for unassigned)
--blocked <bool>true for blocked only, false for unblocked
--parent <id>Filter by parent issue ID (none for top-level)
--created-after <date>Created after date (YYYY-MM-DD)
--created-before <date>Created before date (YYYY-MM-DD)
--due-before <date>Due before date (YYYY-MM-DD)
--due-after <date>Due after date (YYYY-MM-DD)
--sort <field>Sort: id, title, status, priority, type, created, updated, estimate, spent, dueDate (default: id)
--order <order>Sort order: asc, desc (default: asc)
Get
mdp issue get -p <path> --id <id> [--no-include-content]
Update
mdp issue update -p <path> --id <id> [options]
OptionDescription
-t, --title <title>New title
--type <type>New type
-s, --status <status>New status
--priority <priority>New priority
-a, --assignee <assignee>Set assignee (none to clear)
-m, --milestone <milestone>Set milestone (none to clear)
-e, --estimate <estimate>Set estimate (none to clear)
--spent <spent>Set spent (none to clear)
--due-date <date>Set due date (none to clear)
--parent <id>Set parent issue (none to clear)
-l, --labels <labels>Set labels (replaces all)
--add-labels <labels>Add labels
--remove-labels <labels>Remove labels
--blocked-by <ids>Set blockedBy (replaces all)
--add-blocked-by <ids>Add to blockedBy (with cycle detection)
--remove-blocked-by <ids>Remove from blockedBy
--related-to <ids>Set relatedTo (replaces all)
--add-related-to <ids>Add to relatedTo
--remove-related-to <ids>Remove from relatedTo
--add-checklist <items>Add checklist items
--remove-checklist <items>Remove checklist items by text
--check <items>Check items by text
--uncheck <items>Uncheck items by text
-c, --content <content>Replace markdown body
--dry-runPreview without writing
Delete
mdp issue delete -p <path> --id <id> [--dry-run]

Cleans up blockedBy, relatedTo, and parent references in other issues.

Log

Manage log entries on an issue.

mdp issue log add -p <path> --id <id> -b "message" [--author <name>] [--dry-run]
mdp issue log list -p <path> --id <id>
mdp issue log get -p <path> --id <id> --index <n>
mdp issue log update -p <path> --id <id> --index <n> [--author <a>] [-b <body>] [--dry-run]
mdp issue log delete -p <path> --id <id> --index <n> [--dry-run]
Batch create

Create multiple issues at once by piping a JSON array via stdin. Processes items sequentially; continues on error and reports per-item success/failure.

echo '<json-array>' | mdp issue batch-create -p <path> [--dry-run]

Input: JSON array of issue objects. Required field: title. All other fields match issue create options (type, status, priority, labels, assignee, milestone, estimate, spent, dueDate, blockedBy, parent, relatedTo, checklist, description, content, template). Array fields (labels, blockedBy, etc.) accept string[].

Output:

{
  "ok": true,
  "data": {
    "total": 3, "succeeded": 2, "failed": 1,
    "results": [
      { "ok": true, "data": { "id": "ISS-4", "title": "...", "filePath": "..." } },
      { "ok": false, "error": { "code": "INVALID_STATUS", "message": "...", "index": 2 } }
    ]
  }
}

Exits with code 1 if any item failed.

Batch update

Update multiple issues at once by piping a JSON array via stdin. Processes items sequentially; continues on error and reports per-item success/failure. In-memory state is refreshed after each successful update for accurate cycle detection.

echo '<json-array>' | mdp issue batch-update -p <path> [--dry-run]

Input: JSON array of update objects. Required field: id. All other fields match issue update options (title, type, status, priority, labels, addLabels, removeLabels, assignee, milestone, estimate, spent, dueDate, blockedBy, addBlockedBy, removeBlockedBy, parent, relatedTo, addRelatedTo, removeRelatedTo, addChecklist, removeChecklist, check, uncheck, content). Array fields accept string[].

Output: Same envelope format as batch-create, with changes and filePath per item.

Exits with code 1 if any item failed.

Milestones

Create
mdp milestone create -p <path> -t "Title" [options]
OptionDescription
-t, --title <title>Milestone title (required)
-s, --status <status>Initial status (default: first status in config)
--priority <priority>Priority level (default: null)
-l, --labels <labels>Comma-separated labels
--start-date <date>Start date (YYYY-MM-DD)
--due-date <date>Due date (YYYY-MM-DD)
--checklist <items>Comma-separated checklist items
-d, --description <desc>Short description
-c, --content <content>Full markdown body
--template <name>Template name from.mdp/templates/
--dry-runPreview without creating
List
mdp milestone list -p <path> [options]
OptionDescription
-s, --status <statuses>Comma-separated status filter
--priority <priority>Priority filter
-l, --labels <labels>Comma-separated labels filter
--overdue <bool>true for overdue only, false for not overdue
--sort <field>Sort: id, title, status, priority, created, updated, dueDate, completion (default: id)
--order <order>Sort order: asc, desc (default: asc)
Get
mdp milestone get -p <path> --id <id> [--no-include-content]
Update
mdp milestone update -p <path> --id <id> [options]
OptionDescription
-t, --title <title>New title
-s, --status <status>New status
--priority <priority>New priority
--start-date <date>Set start date (none to clear)
--due-date <date>Set due date (none to clear)
-l, --labels <labels>Set labels (replaces all)
--add-labels <labels>Add labels
--remove-labels <labels>Remove labels
--add-checklist <items>Add checklist items
--remove-checklist <items>Remove checklist items by text
--check <items>Check items by text
--uncheck <items>Uncheck items by text
-c, --content <content>Replace markdown body
--dry-runPreview without writing
Delete
mdp milestone delete -p <path> --id <id> [--dry-run]

Clears milestone reference from all assigned issues.

Progress
mdp milestone progress -p <path> --id <id>

Returns completion percentage, issue counts by status, and list of assigned issues.

Log

Manage log entries on a milestone.

mdp milestone log add -p <path> --id <id> -b "message" [--author <name>] [--dry-run]
mdp milestone log list -p <path> --id <id>
mdp milestone log get -p <path> --id <id> --index <n>
mdp milestone log update -p <path> --id <id> --index <n> [--author <a>] [-b <body>] [--dry-run]
mdp milestone log delete -p <path> --id <id> --index <n> [--dry-run]

Search

mdp search -p <path> -q "query text" [--limit <n>]

Searches project, issues, and milestones by text content using BM25 ranking. Returns matched fields with snippets. Use --entity project to search only the project file.

Workflow recommendations

  1. Start with mdp project create -p. --preset software to set up a project
  2. Create issues with mdp issue create — use --type and --labels for organization
  3. For bulk operations, use mdp issue batch-create and mdp issue batch-update with JSON arrays piped via stdin
  4. Move issues through statuses with mdp issue update --id ISS-1 -s "In Progress"
  5. Group work into milestones, then track with mdp milestone progress
  6. Use mdp issue list with filters to find relevant issues

See WORKFLOWS.md for detailed workflow patterns.

References

平台分布

Codex

34.73%
按下载量换算25

Claude

31.6%
按下载量换算23

Cursor

16.78%
按下载量换算12

Gemini CLI

9.56%
按下载量换算7

安全审计

Gen Agent Trust Hub

未通过

Socket

通过

Snyk

通过

权限和风险

只读

该 Skill 主要提供规则、说明或参考内容,本身偏只读;真正读写文件、联网或执行命令仍取决于宿主 Agent 的任务。

安装前确认

本站仅展示第三方公开信息,不托管安装包,不提供自动安装或运行环境。安装前应自行审查源码、依赖和命令行为。来源安全扫描存在提醒或未通过项,安装前请自行复核。目前只有一个公开来源,建议同时核对源仓库。

继续浏览同类 Skills